experiments.sclbridge package

Submodules

experiments.sclbridge.wrapper module

class experiments.sclbridge.wrapper.AvalancheRON(n_inp: int, n_hid: int, dt: float, gamma: float | Tuple[float, float], epsilon: float | Tuple[float, float], rho: float = 0.99, input_scaling: float = 1, reservoir_scaler=0, sparsity=0, device='cpu', n_classes: int = 10)

Bases: RandomizedOscillatorsNetwork

Adapt RandomizedOscillatorsNetwork for Avalanche strategies.

Parameters:

• n_inp – Number of input features per time step.

• n_hid – Reservoir hidden size.

• dt – Reservoir integration step.

• gamma – Damping parameter or sampling range.

• epsilon – Stiffness parameter or sampling range.

• rho – Spectral radius.

• input_scaling – Input scaling factor.

• reservoir_scaler – Scaling for structured reservoirs.

• sparsity – Reservoir sparsity.

• device – Torch device.

• n_classes – Number of output classes.

forward(x: Tensor) → Tensor

Run the reservoir and classify the final hidden state.

Parameters:

x – Input tensor accepted by Avalanche minibatches.

Returns:

Class logits.

Module contents

Avalanche wrappers for ACDS reservoir models.